bugbug/scripts/get_untriaged.py

114 строки
3.5 KiB
Python

# -*- coding: utf-8 -*-
# This Source Code Form is subject to the terms of the Mozilla Public
# License, v. 2.0. If a copy of the MPL was not distributed with this file,
# You can obtain one at http://mozilla.org/MPL/2.0/.
import argparse
import csv
import json
import os
import sys
from datetime import date, datetime, timedelta
# Inject project path
sys.path.append("../")
def fetch_untriaged(args):
from bugbug import bugzilla
today = date.today()
three_months_ago = today - timedelta(days=args.days_back)
# Set bugzilla token and download bugs
bugzilla.set_token(args.token)
bug_ids = bugzilla.get_ids_between(three_months_ago, today)
bugs = bugzilla.get(bug_ids)
# Get untriaged bugs
untriaged_bugs = []
for bug in bugs.values():
for history in bug["history"]:
for change in history["changes"]:
if (
change["field_name"] == "component"
and change["removed"] == "Untriaged"
):
untriaged_bugs.append(bug)
with open("bugs-{}.json".format(datetime.now().strftime("%s")), "w") as f:
json.dump(untriaged_bugs, f)
return untriaged_bugs
def run_untriaged(untriaged_bugs):
from bugbug.models.component import ComponentModel
from bugbug.models.component_nn import ComponentNNModel
models = [
(ComponentModel, "../componentmodel"),
(ComponentNNModel, "../componentnnmodel"),
]
for (model_class, model_file_name) in models:
rows = []
model = model_class.load(model_file_name)
for bug in untriaged_bugs:
p = model.classify(bug, probabilities=True)
url = f'https://bugzilla.mozilla.org/show_bug.cgi?id={bug["id"]}'
classifiable = p[p >= 0.7].size >= 1
classification = ""
expected_component = model.filter_component(
bug["product"], bug["component"]
)
if not expected_component:
print("Skipping bug: {}".format(bug["id"]))
continue
if classifiable:
print("Classifying bug with ID: {}".format(bug["id"]))
classification = model.classify(bug)[0]
print("Classified bug as: {}".format(classification))
else:
print("Not classifiable bug: {}".format(bug["id"]))
correct_prediction = expected_component == classification
rows.append(
[
url,
classifiable,
expected_component,
classification,
correct_prediction,
bug["summary"],
]
)
os.makedirs("sheets", exist_ok=True)
class_name = model.__class__.__name__
timestamp = datetime.utcnow().strftime("%Y-%m-%d")
sheet_name = f"{class_name}-{timestamp}-labels.csv"
with open(os.path.join("sheets", sheet_name), "w") as f:
writer = csv.writer(f)
writer.writerows(rows)
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument("--token", help="Bugzilla API token")
parser.add_argument("--days-back", type=int, help="Days since to fetch bugs")
parser.add_argument("--file_path", help="Days since to fetch bugs")
args = parser.parse_args()
if args.file_path:
with open(args.file_path, "r") as f:
bugs = json.load(f)
else:
bugs = fetch_untriaged(args)
run_untriaged(bugs)